Epstein-Barr Virus Latent Membrane Protein 1 Regulates Host B Cell MicroRNA-155 and Its Target FOXO3a via PI3K p110α Activation

Abstract

Epstein-Barr Virus (EBV) is associated with potentially fatal lymphoproliferations such as post-transplant lymphoproliferative disorder (PTLD), a serious complication of transplantation. The viral mechanisms underlying the development and maintenance of EBV+ B cell lymphomas remain elusive but represent attractive therapeutic targets. EBV modulates the expression of host microRNAs (miRs), non-coding RNAs that regulate gene expression, to promote survival of EBV+ B cell lymphomas. Here, we examined how the primary oncogene of EBV, latent membrane protein 1 (LMP1), regulates host miRs using an established model of inducible LMP1 signaling. LMP1 derived from the B95.8 lab strain or PTLD induced expression of the oncogene miR-155. However, PTLD variant LMP1 lost the ability to upregulate the tumor suppressor miR-193. Small molecule inhibitors (SMI) of p38 MAPK, NF-κB, and PI3K p110α inhibited upregulation of miR-155 by B95.8 LMP1; no individual SMI significantly reduced upregulation of miR-155 by PTLD variant LMP1. miR-155 was significantly elevated in EBV+ B cell lymphoma cell lines and associated exosomes and inversely correlated with expression of the miR-155 target FOXO3a in cell lines. Finally, LMP1 reduced expression of FOXO3a, which was rescued by a PI3K p110α SMI. Our data indicate that tumor variant LMP1 differentially regulates host B cell miR expression, suggesting viral genotype as an important consideration for the treatment of EBV+ B cell lymphomas. Notably, we demonstrate a novel mechanism in which LMP1 supports the regulation of miR-155 and its target FOXO3a in B cells through activation of PI3K p110α. This mechanism expands on the previously established mechanisms by which LMP1 regulates miR-155 and FOXO3a and may represent both rational therapeutic targets and biomarkers for EBV+ B cell lymphomas.

Keywords: Epstein-Barr virus; FOXO3a; PI3K; latent membrane protein 1; miR-155; microRNA.

Figure 1

Tumor variants of LMP1 differentially regulate miR-155 and miR-193b from the B95.8 lab strain of LMP1. Two million EBV− BL41 cells expressing B95.8 lab strain or natural variant NGFR.LMP1 molecules were activated for 12 h prior to RNA isolation with the miRVana miRNA isolation kit. Relative expression of specific microRNAs was determined by quantitative PCR (qPCR) using TaqMan MicroRNA Assays. Target-specific cDNA was generated from 10 ng of total RNA using the TaqMan MicroRNA Reverse Transcription Kit and pre-amplified using the TaqMan PreAmp Master Mix. Finally, qPCR assays were performed using the TaqMan Universal Master Mix II, No AmpERASE UNG. The relative expression of (A) miR-155 and (B) miR-193b was calculated by first normalizing to the endogenous control U47 (ΔC_t) and then to unactivated samples (ΔΔC_t). Fold-induction (2^−ΔΔCt) of each miR is shown. Each point represents an experimental replicate; two different lines expressing B95.8 NGFR.LMP1 were used. ^***p ≤ 0.001, ^****p ≤ 0.0001 by one-way ANOVA with post hoc multiple comparisons to activate B95.8 lab strain NGFR.LMP1.

Figure 2

CTAR2 is essential for regulation of miRs by LMP1, while CTAR1 supports the regulation of miR-155, but not miR-193b. NGFR.LMP1 signaling was activated in 2 × 10⁶ EBV− BL41 cells expressing the following genetically engineered variants of the B95.8 NGFR.LMP1 molecules: wild type (B95.8), CTAR1 mutant (CTAR1mut), CTAR2 mutant (CTAR2mut), or containing both the G212S and S366T mutations commonly found in the natural variants of LMP1 (DM). After RNA isolation, relative expression of (A) miR-155b and (B) miR-193b was determined by quantitative PCR (qPCR) as described in Figure 1. Fold induction (2^−ΔΔCt) of each miR is shown. Each point represents an experimental replicate; two different lines expressing B95.8 NGFR.LMP1 were used. ^**p ≤ 0.01, ^****p ≤ 0.0001 by one-way ANOVA with post hoc multiple comparisons to activate B95.8 lab strain NGFR.LMP1.

Figure 3

The PI3K p110α-specific inhibitor BYL719 inhibits maximal induction of miR-155 by LMP1 in B cells. Two million EBV− BL41 cells expressing B95.8 (A,C) or tumor variant #2 (B,D) NGFR.LMP1 molecules were activated in the presence of the indicated inhibitors for 12 h prior to RNA isolation with the miRVana miRNA isolation kit. Relative expression of specific microRNAs was determined by quantitative PCR (qPCR) as described in Figure 1. Fold induction (2^−ΔΔCt) of each miR is shown. Each point represents an experimental replicate; two different lines expressing B95.8 NGFR.LMP1 were used. ^*p ≤ 0.05, ^**p ≤ 0.01, ^***p ≤ 0.001, ^****p ≤ 0.0001 by one-way ANOVA with post hoc multiple comparisons to activate NGFR.LMP1 in the absence of signaling inhibitors. Inhibitors used: GDC-0941 (GDC), BYL719 (BYL), CAL-101 (CAL), Rapamycin (Rapa), PD98059 (PD), SB203580 (SB), Bay11-7082 (Bay11), cyclosporin A (CSA), and FK506.

Figure 4

PI3K p110α is required for maximal induction of miR-155, but not miR-193b, by LMP1 in B cells. (A–C) Two million EBV− BL41 cells expressing B95.8 NGFR.LMP1 were activated in the presence of the indicated amounts of the PI3K p110α-specific inhibitor BYL719 for 12 h. (A) After lysate generation, Western blots for pAkt (Ser473), Akt, and actin were performed and imaged via the iBright FL1000 imaging system. Representative blots are shown. Densitometry was performed using ImageJ. All values shown were background subtracted, and the ratio of pAkt/Akt was first normalized to actin. Samples were then normalized to the sample where NGFR.LMP1 was activated in the absence of BYL719 to yield Normalized pAkt/Akt. Each point represents an experimental replicate. ^***p ≤ 0.001, ^****p ≤ 0.0001 by one-way ANOVA with post hoc multiple comparisons activated NGFR.LMP1 in the absence of signaling inhibitor; no other significant differences were observed. (B,C) After RNA isolation with the miRVana miRNA isolation kit, relative expression of (B) miR-155 and (C) miR-193b was determined by quantitative PCR (qPCR) as described in Figure 1. Fold induction (2^−ΔΔCt) of each miR is shown. Each point represents an experimental replicate. ^*p ≤ 0.05, ^***p ≤ 0.001 by one-way ANOVA with post hoc multiple comparisons to activate NGFR.LMP1 in the absence of signaling inhibitor.

Figure 5

miR-155 targets FOXO3a and PI3K p85α are downregulated in EBV+ B cell lymphomas. Lysates from EBV+ and EBV− B cell lymphomas were generated with RIPA buffer supplemented with 1× Halt Protease and Phosphatase Inhibitors and 1 mM sodium orthovanadate and quantified using the Pierce 660 nm Protein Assay. About 30 μg of protein was loaded on a 4–20% tris-glycine gel and subsequently transferred to a nitrocellulose membrane. Western blots for the indicated proteins were performed as per manufacturer’s instructions and imaged via the Azure c300 digital imager. Representative blots are shown in (A). Densitometry was performed using ImageJ, and all values were background subtracted and normalized to the indicated loading control (B–E). Each point represents an experimental replicate. ^*p ≤ 0.05, ^**p ≤ 0.01, ^****p ≤ 0.0001 by linear mixed effects model.

Figure 6

miR-155 and miR-193b are increased in EBV+ B cell lymphomas and their associated exosomes. Exosomes were isolated from the supernatants of EBV+ and EBV− B cell lymphoma using the ExoQuick-TC exosome isolation kit. RNA from cells (A,C) and exosomes (B,D) was isolated with the miRVana miRNA isolation kit. Expression of specific microRNAs was determined by quantitative PCR (qPCR) as described in Figure 1. For absolute quantification of these miRs, a standard curve was also performed using Universal Human miRNA Reference RNA. Each point represents unique EBV+ or EBV− B lymphoma cell line.

Figure 7

LMP1 downregulates expression of the miR-155 target FOXO3a via PI3K p110α in B cells. (A–C) Four million EBV- BL41 cells expressing B95.8 NGFR.LMP1 were treated as indicated for 16 h prior to lysis. Lysates were generated using phospholysis buffer supplemented with 1× Halt Protease and Phosphatase Inhibitors and 1 mM sodium orthovanadate and quantified using the Pierce 660 nm Protein Assay. Lysates were loaded on a 4–20% tris-glycine gel and subsequently transferred to a nitrocellulose membrane. Western blots for the indicated proteins were performed as per manufacturer’s instructions and imaged via the iBright FL1000 imaging system. Representative blots are shown. Densitometry was performed using ImageJ, and all values shown were background subtracted and normalized to actin. Each point represents an experimental replicate. ^***p ≤ 0.001, ^****p ≤ 0.0001 by one-way ANOVA with post hoc multiple comparisons. (D) Proposed model by which B95.8 LMP1 supports the regulation of miR-155 and FOXO3a through PI3K p110α in B cells (dark gray). As shown, our data also confirm previous findings regarding the involvement of NF-κB and p38 in the upregulation of miR-155 by LMP1 (light gray). Created with BioRender.com.